Volume replication problem with different versions?
Hi,
I've here two storage systems with Open-E DSS. The hardware is 100% identical. The only difference are the versions of Open-E:
System 1: 5.0.DB49000000.3278
System 2: 5.0up60.7102.3511 32bit
System 1 has some volumes online and can't be stopped.
Volume replication via a single backbone cable is activated. On the consoles, the systems can be ping each other via that backbone link (where unique, non-routed IPs are used)
I created a 16GB iSCSI volumes with replication flag on each system and marked the one on the newer system 2 as source, the other one on the older system 1 as target for volume replication.
Problem: When I scan on system 2 for the corresponding replication volume of system 1, no volume is found => no replication task can be created.
Remark: I've here another twin system with lots of running replications tasks, constructed in the same manner. It works well, but both of these twins have the same (old) version 5.0.DB49000000.3278 ...
Maybe a problem of replication between the different versions of system 2 and 1? Should not be, but I've looked around and couldn't find a misconfiguration.
As concerns volume replication, you need to make sure that the DRBD daemon version is the same on both Open-E machines (you can check this by downloading the system logs and examining the file drbd.log). If the version is the same, it should work.
things are working now. Fortunately, we had a massive VMware storage problem today, thus we were able to update the older Open-E to the same version as the other one (because lots of things didn't work anymore, so we could shut down the complete system cluster).
Immediately, the replication volumes have been displayed as expected.
=> It doesn't suffice that the DRBD versions are the same. The main DSS version must be identical!
(It is not clear to me how one could migrate to a newer version without downtime under that restriction. This is a pity. Open-E should present a solution for this case).